[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: How to debug strange value changes of a variable?
From: |
Marcin Borkowski |
Subject: |
Re: How to debug strange value changes of a variable? |
Date: |
Thu, 01 Jan 2015 01:23:26 +0100 |
On 2014-12-31, at 22:09, Michael Heerdegen <michael_heerdegen@web.de> wrote:
> Marcin Borkowski <mbork@wmi.amu.edu.pl> writes:
>
>> Now somehow mu4e (I know it is to blame, since in stock message-mode,
>> with emacs -Q, and without mu4e, everything works as expected!) messes
>> around with the value of this variable /after the first increment/ (or
>> so it seems: it looks like that after some initialization - including
>> incf'ing that variable from its inital value of -1 - it somehow comes
>> back to its original (default) value). I'd like to check what function
>> and when changes its value. Is there any way to use Edebug (or anything
>> else) for this?
>
> Are you sure you reference the variable's value with the same buffer
> current? That's important of course.
Yes, this I did check.
> Also note that changing a buffer's major mode removes all buffer local
> bindings in the current buffer. That could maybe happen in your case.
Thank you, thank you, thank you, thank you so much! I'd *never* think
of that (I didn't even know that). That was it! The first time my
function was called, the buffer was in fundamental mode; all subsequent
times, it was in mu4e-compose-mode. That should explain everything.
Now the question is: what to do. I could dive into mu4e sources (which
I'm just a bit too lazy to do now). I could just start
mu4e-compose-mode myself in my function and see what happens (I'll try
that). Other than that, I don't have many ideas.
>> I know about edebug-set-global-break-condition, but this doesn't help a
>> lot: it can stop when the variable has some value, but does not tell me
>> /which/ piece of code changed it.
>
> Indeed, that is not much helpful in this case, unless you instrument all
> code that could be responsible, but that may slow down things
> drastically.
What's more, some things might be done by functions written in C...
> Did you define the variable we speak about by yourself? If you did,
> normally only your own code should change the variables binding.
Yes.
>> (A similar question is to find out what calls such-and-such function.
>> This one seems easier: I can just instrument the said function and then
>> press `d' in Edebug to see the backtrace. Am I right?)
>
> Yes, you could. But for such questions, the ordinary debugger is IMHO
> much simpler and more helpful (`debug-on-entry'). Load the elisp source
> files before using it to avoid seeing byte code in the debugger output.
Thanks for the tip!
> Michael.
Thank you again and a happy new year!
--
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Faculty of Mathematics and Computer Science
Adam Mickiewicz University